In this interview, University of Chicago economist and Stone Center Affiliated Scholar Steven Durlauf discusses his working paper on the Great Gatsby Curve and how sociology has influenced his work on poverty, inequality, and economic growth.
You’ve integrated sociological ideas into your research throughout your career. As an economist, what have you learned from sociology?
Durlauf: Sociology has influenced my research in many ways, especially as my work evolved in the direction of studying inequality and intergenerational mobility. As I focused on phenomena, as opposed to methods, it was natural to look at the limitations of how economists conventionally modeled inequality and to see how one can more comprehensively understand such complex phenomena. Much of the study of intergenerational mobility comes down to asking how the incomes of parents delimit or create opportunities for their children, and I have found it impossible to ignore ideas that originate in sociology if one is going to understand this.
The neoclassical economic models of mobility emphasize how families have direct influences on children through investments of time and resources. But there’s a whole host of other mechanisms that suggest that families matter because of the social influences they lead children to experience. Communities defined by geography, ethnicity, and class all have powerful influences. One would say the same thing about schools. Once you take that perspective, you move into ideas that are of ancient lineage in sociology. In my study of these types of questions, it was almost inevitable that sociological ideas would come to the forefront.
The second thing I would say is there’s a dialectic in how one does research. Sociologists, of course, think about questions differently than economists. It’s not just a matter of what variables they assume affect individuals; it’s the way theories are formulated, and the ways that theory and empirics interact, and drawing conclusions about explanations and policy. My own orientation tends to be more theoretical. I think about formal theories of behavior and ask: What are the restrictions these theories place on data? I have always tried to construct formal theories that have statistical analogs. But what goes into these models?
To give a very specific example: sociology has a set of statistical models in which neighborhoods are background contexts against which families affect children. Economists tend to think additively; therefore the effects of neighborhoods and parents on children are modeled as additive. That sounds like a technical detail, but it’s actually a substantive difference. The traditional sociologist’s way of treating neighborhoods as a background leads to different statistical models, and what are called hierarchical models in sociology are not popular in economics. It turns out that the conditions for untangling neighborhood and family influences in statistical models are very different for hierarchical models than for economists’ preferred so-called linear-in-mean models, something I have shown in work with Lawrence Blume and William Brock.
You’ve said that working on your Great Gatsby Curve working paper reinforced to you how much needs to be done in order for economic models to catch up to the richness of a sociological perspective. What changes would you like to see?
Durlauf: In thinking about intergenerational mobility, the first question is: Mobility of what? Economists traditionally look at income, while sociology looks at occupational mobility. In some sense that’s no surprise; after all, sociology emerged in the 19th century in response to issues of class. So the objectives of economics and sociology in studying mobility have been traditionally somewhat different. Working on the Gatsby paper reinforced in my mind that these need to be integrated, and that there is a better way to think about the dynamics of inequality.
The sociology literature is far beyond the economics literature in understanding occupations, in particular with respect to the idea that there’s something about the work that people do that transmits from parents to children. That’s what I would want to focus on — this more comprehensive view that parental income, parental education, and parental occupation are all mechanisms by which parents shape their children’s opportunities. Now, this can be pushed further to include the intergenerational creation of cultural capital. In other words, how parents affect the aspirations, norms, and tacit knowledge of their children all are major missing pieces in traditional economics approaches to mobility.
In my own research, I am working to develop such comprehensive models of this process by which parents influence children. There’s a complicated set of things that characterizes parents: their own income trajectories, their family structure trajectories, their education levels and backgrounds, the locations where they reside — all of these are influences and I want to parse out the respective roles further and see how they interact. I have been fortunate to begin collaborations with two brilliant senior sociologists in this work, Felix Elwert at Wisconsin and Xi Song at Penn, as well as some remarkable junior scholars, David McMillon in economics at Emory and Gueyon Kim in economics at UC Santa Cruz.
What did you conclude in your study of the theoretical and empirical work on the Great Gatsby Curve?
Durlauf: First, I am increasingly convinced that one should not think of inequality and mobility as distinct, but rather as two dimensions of the complex stochastic process that describes family income dynamics. In other words, the Gatsby Curve should be thought of as describing the two statistics as proxies for two features of a complicated mathematical process. With Aleksandra Lukina, an extraordinary postdoc at Harris, and my longtime coauthor Lawrence Blume at Cornell, we are formulating ways to operationalize this idea that involve new measures of intergenerational mobility. One conclusion is that new work should go beyond the Gatsby Curve to consider more complex notions of cross-sectional inequality and intergenerational persistence. This conclusion applies to both the mathematics of the Gatsby Curve and the substance of inequality and mobility. By the former, I mean that there are deep logical links between mobility and inequality that researchers have only begun to explore. By the latter, I mean, for example, that inequality has many facets, and the links to mobility will depend on which facets are examined.
Second, the work reinforced my belief that segregation of various types are an essential mechanism in understanding how changes in cross-sectional inequality affect persistence in status across generations. The more inequality there is, the more segregation there is in society. In other words, if one thinks about residential segregation by income, or thinks about test score segregation by colleges, or skill segregation by firms, in each case levels of segregation are higher when there’s more inequality. And the reason is that the incentives for individuals to isolate themselves shift, given the amount of inequality and the gaps between individuals. So in my view of the dynamics, a general story has these features: inequality begets segregation, segregation begets higher disparities in the experiences of affluent and less affluent children, and higher disparities beget greater persistence of socioeconomic status.
A number of years ago I coined the phrase “the memberships theory of inequality” to characterize a perspective that says that that one of the lenses through which to see inequality is via the many different memberships that people have, and these memberships have causal influences. A fundamental membership for children is, of course, their parents. But others are deeply consequential as well. We’re all members of neighborhoods, communities, ethnic groups, genders, cultures, and religions, all of which simultaneously influence child development. As individuals grow, there are other memberships that emerge: what schools the person attends, what jobs they have, what organizations they belong to. These all interact. And in my own work, the focus has been on income and types of segregation as the mechanisms that map greater inequality to greater persistence of socioeconomic status.
Steven Neil Durlauf is the Steans Professor in Educational Policy at the University of Chicago Harris School of Public Policy.
Read the Paper: The Great Gatsby Curve, by Steven N. Durlauf, Andros Kourtellos, and Chih Ming Tan